Experience has shown that airbags work best in combination with seat belts and other safety systems. Although airbags contribute to the overall safety of occupants of an automobile, they can present a danger to a vehicle occupant who is positioned too close to an airbag when it deploys. This condition, where the vehicle occupant is positioned so that airbag deployment might be dangerous, is referred to as the vehicle occupant being “out of position.” Various systems have been developed to detect an “out of position” vehicle occupant. Sensor systems designed to detect the vehicle occupant's position often require constant monitoring so that in the event of a crash the vehicle occupant's position is known. Sensor systems designed to detect the position of the vehicle occupant have been proposed based on ultrasound, optical, or capacitance sensors.
Constant monitoring of sensors, which may have high data rates, requires the design of algorithms which can reduce sensor data to a single condition or a limited number of data conditions which are used in an airbag deployment decision to prevent airbag deployment or for a duel stage airbag to select the level of deployment. Maintaining data integrity between the non-crash positional data, and positional data needed during airbag deployment is complicated by the noisy environment produced by a crash. Dealing with data integrity issues requires increased processor capabilities and algorithm development, which also requires additional testing.
Prior art approaches attempt to determine, based on various sensors, the distance between the airbag and the passenger before the airbag is deployed. In many instances, the vehicle occupant will not be too close to the airbag at the time the decision to deploy the airbag is made, but, because of the rate at which the vehicle occupant is approaching the airbag, the vehicle occupant will be too close when the airbag is actually deploying. To handle these situations, more sophisticated sensors and algorithms are needed in order to attempt to predict the vehicle occupant's position when the airbag is actually deployed or nearly completely deployed. In other words, the ideal airbag deployment system functions such that the airbag deploys fully or nearly fully before the vehicle occupant engages the airbag. Existing systems inhibit airbag deployment when, based on various sensors and algorithms, it is determined that, because of the position of the vehicle occupant, the bag is more likely to harm than to benefit the vehicle occupant.
Successfully creating a sensor and algorithm system is complicated because there is usually very little delay between the decision to deploy and actual deployment. This is so because the maximum benefit from an airbag is achieved by early deployment, and at the same time, more time before deployment maximizes the information available to determine whether deployment is necessary. The desire to maximize effective deployment of the airbag while minimizing unnecessary deployment creates a tension between waiting for more information and deploying immediately. Therefore, once sufficient information is available, deployment typically follows nearly immediately.
A system which employs vehicle occupant position sensors and algorithms must be able to supply at all times an indication of whether airbag deployment should be inhibited so that the inhibit decision can be applied whenever the airbag deployment decision occurs. This means the sensors and algorithms used to develop the vehicle occupant position inhibit signal, cannot be optimized to deal with a specific time frame in which the actual deployment decision is made. The end result is that such algorithms may be less accurate than desired because they must predict events relatively far in the future—perhaps tens of milliseconds.
One known type of sensor shown in EP 0990567A1, employs a plurality of tapes which extend between the front of the airbag and a tape dispensing cartridge mounted on the airbag housing. Tape extraction sensors within the cartridge monitor the rate at which tape is withdrawn from the cartridge and thus can detect airbag impact with a vehicle occupant by a decrease in airbag velocity. This type of sensor which can monitor the way an airbag is actually deploying solves the problem of predicting whether a vehicle occupant will be out of position at time of airbag deployment. In this arrangement the airbag is deployed, and if it encounters a vehicle occupant before it has reached a certain stage of deployment the airbag is vented which effectively removes the airbag. Several tapes and tape dispensing cartridges are used to monitor different portions of the bag so that if any portion of the bag contacts a vehicle occupant, the fact of contact can be detected and the bag vented to prevent injury to the out-of-position occupant. To be practical, this type of sensor—which monitors actual deployment—needs simple but robust techniques for monitoring the rate at which tape is withdrawn from the cartridge.